Estimating the rapid haemodynamic effects of passive leg raising in critically ill patients using bioreactance.

Abstract

Rapid detection of changes in cardiac index (CI) in real time using minimally invasive monitors may be of clinical benefit. We tested whether the Starling-SV bioreactance device, which averages CI over a short 8 s period, could assess the effects of passive leg raising (PLR), a clinical test that is recommended to assess fluid responsiveness during septic shock. In 32 critically ill patients, we measured CI by transpulmonary thermodilution (PiCCO2, CI_td), pulse contour analysis (PiCCO2, CI_Pulse), and the Starling-SV device (CI_Starling) at baseline. CI_Pulse and CI_Starling were measured again at the end of a PLR test. In the 13 patients with a positive PLR test, CI_td, CI_Pulse, and CI_Starling were measured before and after a 500 ml saline infusion. The primary outcome was relative changes from baseline measurements in CI_td, CI_Pulse, and CI_Starling. Secondary outcomes compared absolute values measured by each method. Relative changes in CI_Pulse and CI_td were significantly correlated (r=0.82; n=45; P<0.001), with an 89% concordance rate (n=45 paired measurements). Relative changes in CI_Starling and CI_td were also significantly correlated (r=0.59; n=45; P<0.001) with a 78% concordance rate. For absolute measures of CI (n=77 paired measurements), the bias between CI_Pulse and CI_td was 0.01 L min^-1 m^-2 (limits of agreement, -0.49 and 0.51 L min^-1 m^-2; 15% percentage error). Bias between CI_Starling and CI_td was 0.03 L min^-1 m^-2 (limits of agreement, -1.61 and 1.67 L min^-1 m^-2; 48% percentage error). In critically ill patients, a non-invasive bioreactance device with a shorter averaging period assessed a passive leg raising test with reasonable accuracy.